13th World Conference on Earthquake Engineering

Vancouver, B.C., Canada

August 1-6, 2004
Paper No. 2025

CURRENT DEVELOPMENTS AND FUTURE PROSPECTS OF THE

EUROPEAN CODE FOR SEISMIC DESIGN AND REHABILITATION OF
BUILDINGS: EUROCODE 8
Michael N. FARDIS 1

SUMMARY

The technical contents of the recently approved Part 1 of Eurocode 8: General rules, seismic actions, rules
for buildings [1] and of draft Part 3: Assessment and retrofitting of buildings [2] are reviewed, along with
the procedure for the upcoming enforcement of the 1st generation of European structural design standards.
Possible future technical developments towards the 2nd generation of these parts of EC8 are highlighted.

INTRODUCTION: TOWARDS THE 1ST GENERATION OF EN-EUROCODES

Conversion of about 60 ENV-Eurocodes - published by the European Standards Organisation (CEN) in

the early to mid-90s as European prestandards (ENVs) - into the 1st generation of European Structural
Standards (EN) started in 1998. It entails a thorough revision of the ENVs on the basis of comments of
National Standard Bodies and of recent scientific/technical developments. In the set of 10 Eurocodes, two
cover the basis of structural design and the loadings (“actions”), one covers geotechnical and foundation
design and five cover aspects specific to concrete, steel, composite (steel-concrete), timber, masonry or
aluminum construction. Instead of distributing seismic design aspects to the Eurocodes on loadings,
materials or geotechnical design, all aspects of seismic design are covered by Eurocode 8: “EN1998:
Design of Structures for Earthquake Resistance” (EC8). This is for the convenience of countries with very
low seismicity, as it gives them the option not to apply Eurocode 8 at all.

Eurocode 8 has 6 parts: Part 1 (EN1998-1): “General rules, seismic actions, rules for buildings”; Part 2
(EN1998-2): Bridges; Part 3 (EN1998-3): “Assessment and retrofitting of buildings”; Part 4 (EN1998-4):
“Silos, tanks, pipelines”; Part 5 (EN1998-5): “Foundations, retaining structures, geotechnical aspects”;
Part 6 (EN1998-6): “Towers, masts, chimneys”. In March 2004 Parts 1 and 5 were officially approved as
European Standards (ENs) by the 26 National Standard Bodies in CEN. It is expected that in early 2005
Parts 3 and 2 (in that order) will be approved as ENs, followed by Parts 4 and 6 around the end of 2005.

The European Commission considers the EN-Eurocodes as tools for the integration of the construction
market and all related services in the European Union (EU) and the enhancement of competitiveness of
European designers, contractors, consultants and construction products within and outside the EU. To this
end, all parts of the EN-Eurocodes are developed to be fully consistent, user-friendly and seamlessly

1
Dept. of Civil Engineering, Univ. of Patras, P.O. Box 1424, GR26500, GREECE, fardis@upatras.gr
integrated and to cover fully all types of civil engineering works, except nuclear power plants, off-shore
structures and long-span (cable-stayed or suspension) bridges. They will be the basis of technical
specifications for public works tenders or public procurements in EU countries and the related
engineering services. Nonetheless, at least for the 1st generation of EN-Eurocodes, specification of
national regulations as acceptable alternatives in such tenders will also be allowed, as long EU Member
States retain them in parallel use with EN-Eurocodes. The European Commission will also expect
Member States to promote the EN-Eurocodes in private contracts, instead of any national regulations that
may retain in use and will urge them to phase out national regulations in the medium term.

Member States have the right to set the level of safety, serviceability and durability provided by structures
built on their territory, taking into account aspects of economy. To allow them to exercise this right
without sacrificing harmonisation of structural design codes at the European level and to accommodate
geographic, climatic, etc. differences (including seismotectonic differences), “Nationally Determined
Parameters” (NDPs) have been introduced in the EN-Eurocodes. NDPs are also used wherever consensus
cannot be reached on aspects not related to safety, durability, serviceability or economy. The NDPs are:
- key parameters controlling safety, durability, serviceability or economy (safety factors, the mean
return period of the design seismic action, etc.); the normative text of the Eurocode may provide a
range of permissible values for them and will normally recommend a value in a non-normative note.
- technical classes (e.g. classes of ductility, classes of importance depending on occupancy or use).
- alternative procedures or methods (e.g. models of calculation).
Alternative NDP-classes and NDP-procedures/methods are identified and fully described in the normative
text. A non-normative note may recommend one of the alternative NDP-classes or procedures/methods.

National choice regarding the NDPs will be exercised through the National Annex, to be published by
each country as integral part of the national version of the EN-Eurocode. To minimize variation within the
EU, the European Commission urges Member States to adopt for the NDPs the choice recommended in
the notes. A National Annex may also contain country-specific data (a seismic zonation map, spectral
shapes for the types of soil profiles provided in EC8, etc.) that are also NDPs. The decision to adopt
nationally or not an Informative Annex of the EN-Eurocode may also be made there. If the National
Annex does not exercise national choice for some NDPs, the choice will be left to the designer, taking into
account the conditions of the project and other national provisions. What a National Annex cannot do is
modify any EN-Eurocode rule or replace it with another – e.g. national – rule. Although discouraged, such
a deviation from the EN-Eurocodes is allowed in national regulations other than the National Annexes.
However, if the design includes such deviations, it will not be called “design according to EN-
Eurocodes”, as this implies compliance with all EN-Eurocode provisions, including the national choices
for the NDPs. To help the user apply the EN, a National Annex may include reference to national
documents for supplementary information, non-contradictory to the rules of the Eurocode. Examples for
EC8 may be:
- acceptable values of member plastic hinge or chord rotations for the non-collapse performance level,
(as a function of the geometry and the mechanical properties of the member) for design on the basis
of nonlinear analysis, with direct verification of deformations, rather than of resistances.
- mechanical models for the (masonry) infill panels in infilled frames, for use whenever explicit
modelling of such infills is desired or needed.
A Member State does not need to introduce a National Annex for a Eurocode part considered not relevant
for its territory. This may be the case for EC8 in non-seismic or very low seismicity EU countries.

In less than one year after formal approval of an EN-Eurocode, EU Member States will have to translate it
into their national language, or adopt one of the three CEN official versions (English, German or French).
The deadline for publishing the National Annex – including the national choice for the NDPs – is two
years after formal approval of the EN-Eurocode. Member States are expected to calibrate the NDPs during
these two years, so that, for the target safety level, structures designed according to the national version of
the EN-Eurocodes will not cost significantly more than if designed with current National Standards. For
transparency within the internal market, EU Member States will have to make clearly known to Eurocode
users and any other interested party their national choice for the NDPs. Soon after the publication of all
National Annexes, the European Commission intends to review the national choices for NDPs. Depending
on the outcome, the European Commission may request EU Member States to accept a unique choice for
as many NDPs as possible, especially for those ones for which national choices are not very different
and/or the final design turns out to be not so sensitive to the national choices of the NDP.

EN Eurocodes will be applied in packages, each package referring to a specific combination of type of
structure and construction material (e.g., concrete buildings, steel silos, tanks or pipelines, composite
bridges, etc.). For self-sufficiency, each package will include EN1990: “Basis of Structural Design”,
EN1997: “Geotechnical Design” and Parts 1 and 5 of Eurocode 8, as well as the appropriate parts of
EN1991: “Actions on Structures”. Part 3 of EC8 will be included in all packages for buildings, Part 2 in
all packages for bridges, Part 4 in the packages for “concrete liquid retaining and containment structures”
and for “steel silos, tanks and pipelines” and Part 4 in the package for “Steel towers and masts”.

Member States must withdraw national standards that compete or conflict with any EN-Eurocode part in a
package, five years after the last EN-Eurocode in the package is published by CEN (i.e. three years after
the deadline for publication of the National Annex of this last Eurocode part). This will be the beginning
of the exclusive use of the EN-Eurocodes in the package as structural design standards in the EU.

A maintenance group will be set up for each EN-Eurocode after its official approval. Its task will be:
- to ensure consistency and co-ordination of on-going work on other draft ENs,
- to collect comments, provide explanations and answers to questions,
- to prepare the next phase of revision and updating of the EN, on the basis of feedback from its
application and of scientific and technical developments.
The revision of the 1st generation of EN-Eurocodes and the development of the next one will be a natural
continuation of the work of maintenance groups. To prevent confusion, a second generation EN will not
be published by CEN, before the package of first generation of EN-Eurocodes it belongs to becomes the
exclusive set of European structural design standards and conflicting national standards are withdrawn.

The paper focuses on the parts of EC8 devoted to the design of new buildings - Part 1: EN1998-1: 2004
[1] - or the assessment and retrofitting of existing ones - Part 3: draft prEN1998-3 [2]. Starting with the
Introduction the paper summarizes recent developments and presents short- or mid-term prospects, either
as these are planned, or as they are foreseen/sensed by the author on the basis of his experience in charge
of the conversion of the six parts of Eurocode 8 from prestandards (ENVs) to European Standards (ENs).

PART 1 OF THE EN-EUROCODE 8 FOR SEISMIC DESIGN OF BUILDINGS

Introduction
Although its main object is buildings, Part 1 of the EN-Eurocode 8 includes also the general provisions for
the other parts of EC8 to build on: Performance requirements, seismic action, analysis procedures, general
concepts and rules applicable to structures beyond buildings. It covers in separate chapters concrete, steel,
composite (steel-concrete), timber and masonry buildings, buildings with base isolation.

Performance Objectives: Seismic action and associated performance requirements

Part 1 of EC8 provides for a two-level seismic design, with the following explicit performance objective:
- Protection of life under a rare seismic action, by prevention of collapse of the structure or parts
 thereof and preservation of structural integrity and of residual load capacity.
- Limited property loss in a frequent earthquake, via limitation of structural and non-structural damage.
The no-local-collapse performance level is achieved by proportioning and detailing structural elements for
a combination of strength and ductility providings a safety factor between 1.5 and 2 against substantial
loss of lateral load resistance. The damage limitation performance level is pursued by limiting the overall
deformations (lateral displacements) of the system to levels acceptable for the integrity of all its parts
(including non-structural ones) and through non-engineered measures for the integrity of (masonry) infills.

Although not explicitly stated, a third objective is the prevention of global collapse in an extremely strong
earthquake, of the order of the “Maximum Considered Earthquake” (MCE) of US codes (although it is
recognized that repair may be unfeasible or economically prohibitive and that the damaged structure may
collapse in a strong aftershock). Resistance to extremely strong seismic actions is pursued by control of
the inelastic response mechanism through systematic and across-the-board application of capacity design.

Within the philosophy of national authority on issues of safety and economy, hazard levels corresponding
to the two performance levels above are left for national determination. For structures of ordinary
importance the recommendation in Part 1 of EC8 is for:
- A seismic action for (local) collapse prevention with 10% exceedance probability in 50 years (mean
return period: 475 years). This is the “design” seismic action; the “design” seismic action of
structures of ordinary importance on rock is termed “reference” seismic action.
- A 10% in 10 years “serviceability” action for damage limitation (mean return period: 95 years).

Enhanced performance of essential or large occupancy facilities is achieved not by upgrading the
performance level for given earthquake level, as in US codes, but by modifying the hazard level for which
the performance level is pursued. For essential or large occupancy structures the seismic action at both
performance levels should be increased so that its exceedance probability in 50 or 10 years, respectively,
is lower than 10%. At the collapse prevention level the recommended value of the NDP-importance factor
γI (to be applied to the reference seismic action) is 1.4 or 1.2 for essential or large occupancy buildings,
respectively. A γI-value of 0.8 is recommended for buildings of reduced importance for public safety.

It is meant that the same spectral shape will be used for the seismic action for damage limitation as for
collapse prevention, with a single multiplicative factor reflecting the difference in hazard level. The value
of this factor should reflect national choice regarding protection of property, but also the regional seismo-
tectonic environment. Values of 0.4 and 0.5 are recommended for this NDP-conversion factor, giving at
the end approximately the same property protection for ordinary and large-occupancy buildings, less
property protection for buildings of low importance (by 15-20% at the level of the seismic action) and
higher property protection for essential facilities (by 15-20% at the level of the seismic action), possibly
allowing them to operate during or immediately after a frequent event. In buildings, and unless most of the
lateral force resistance will be provided by walls, the size of members will be controlled by the limitation
of story drift ratio under the 10% in 10 years “serviceability” seismic action, calculated on the basis of the
equal-displacement rule. The drift limit is 0.5% if the non-structural elements are brittle and attached to
the structure, 0.75% if they are ductile, or 1% if they are not forced to follow structural deformations or do
not exist at all. The 1% drift limit protects also structural members from extensive inelastic deformations
under the “serviceability” earthquake. As the National Annex will set the level of “serviceability”
earthquake, it will also determine to which extent these limits will control member dimensions. With the
EC8-recommended values of 0.5x0.8=0.4 to 0.4x1.4=0.56 for the ratio of the “serviceability” to the
“design” seismic action, the EC8 limits are 2 to 3 times stricter than in US codes.

The elastic response spectrum of the “design” seismic action is anchored to the “reference” ground
acceleration on rock, αgR, to be mapped in national zoning maps. The spectrum includes regions where
spectral acceleration, pseudovelocity or displacement is constant. The extent of each of these regions and
the corresponding spectral amplitudes are taken to depend on the top 30m of the ground (which may be
rock or various types of soil) and on the geologic features underneath. The standard ground types are:
- Type A: rock, with a lower limit on the average shear wave velocity in the top 30m, vs,30, of 800m/s;
- Type B: very dense sand or gravel, or very stiff clay, with vs,30 from 360 to 800m/sec;
- Type C: medium-dense sand or gravel, or stiff clay, with vs,30 from 180m/sec to 360m/sec;
- Type D: loose-to-medium sand or gravel, or soft-to-firm clay, with vs,30 less than 180m/sec;
- Type E: 5m to 20m thick soils with vs,30 less than 360m/sec, underlain by rock.
Special studies are needed for the definition of the spectrum over soils containing at least 10m of soft
clay/silt with high plasticity index and high water content, or over liquefiable soils, sensitive clays, etc.

The entire elastic spectrum is anchored to the mapped “reference” acceleration on rock multiplied by: a)
the importance factor γI; b) a (NDP) factor S≥1 reflecting the effect of ground conditions on spectral
values; and c) a correction factor for damping ζ other than 5%, equal to 10/ (5 + ζ ) .

Figure 1. Elastic spectra of Type 1 (left) and 2 (right) recommended for the 5 standard ground types

The definition of spectral shapes for the standard ground types is left to National Annexes, depending on
the magnitude of earthquakes contributing most to the hazard. In principle this allows a National Annex to
introduce spectral amplification factors over each of the 3 regions of the spectrum – constant acceleration,
pseudovelocity or displacement – which decrease with the absolute magnitude of the spectral value, due to
the soil nonlinearity effect. This is the approach already taken in the USGS Seismic Hazard Maps to be
used with all recent nationally applicable US documents. Instead of spectral amplification factors that
decrease with increasing design acceleration (spectral or ground) as in recent US codes, the (non-binding)
recommendation in Part 1 of EC8 is for two constant spectral shapes:
- Type 1 for moderate to large magnitude earthquakes;
- Type 2 for low magnitude ones (e.g. with surface magnitude less than 5.5) at close distances
(producing over soft soil a motion rich in high frequencies).
Figure 1 shows the two recommended spectra types for ζ=5%, γI=1 and a PGA on rock of 1.0g. These
normalized spectra were developed from a fairly large database of available records from Southern
Europe and the Mediterranean area. They are believed to be currently more representative of the seismic
hazard of that region than the more refined description of elastic spectra in recent US codes. Nonetheless,
they are clearly not representative of intermediate depth earthquakes from the Vrancea region in Romania
or the Aegean. Noteworthy is the rapid decay of the spectra in Figure 1 for long periods; the importance of
this decay is limited though by the lower limit of 20% of acceleration at zero period recommended for
design spectral accelerations (i.e. after division by the behavior factor, q).

The 2nd generation of the EN-Eurocode 8 will most likely see a redefinition of the standard ground types,
with possible subdivision of some of the current types, e.g. of C. It may also see seismic hazard maps for
the entire European area, seamless at national boundaries. The maps should be in terms of 5%-damped
elastic spectral values (for the constant acceleration, pseudovelocity or displacement regions), given at
various different mean return periods, to let each European country exercise the right to choose the
seismic action for the various performance levels foreseen in the different parts of EC8 (currently two in
Part 1 or three in Part 3). It remains to be seen whether such maps will be included - along with the
definition of the seismic action in terms of the response spectrum - in a special part of EN1991 (“Actions
on Structures”), like the parts on snow or wind loads including the European snow and wind maps, or
whether EC8 will remain the sole Eurocode covering aspects of seismic design, including the loading.

Design of buildings for controlled inelastic response and energy dissipation

The standard procedure in the 1st generation of the EN-Eurocode 8 is force-based design on the basis of
the results of linear analysis for the 5%-damped elastic spectrum reduced by the “behavior factor” q.

EC8 allows neglecting the seismic action, if it is considered so low that design for the other actions
provides also the necessary earthquake resistance. It is recommended (as NDP) that seismic design is not
needed if the reference ground acceleration at the ground surface, SagR, times the importance factor, γI, is
less than 0.05g. EC8 allows also providing earthquake resistance by just proportioning members
according to the other Eurocodes, for internal forces from an analysis for the seismic action with forces
reduced by a behavior factor (q) of 1.5 (or 2 in some cases of steel or composite buildings) considered
available due to overstrength, without any measures for local or global ductility. For non-base-isolated
structures this simplification of design is recommended only for low seismicity regions. Although the
designation of a region as one of low seismicity is left to the National Authorities as a NDP, a threshold
value of 0.1g is recommended for the product of γI times the design ground acceleration at the ground
surface (γISαgR). Building structures designed with just the material Eurocodes and a q value of 1.5 (or 2
steel or composite) are termed “low-dissipative” or of Ductility Class (DC) low (L).

The majority of structures designed with EC8 will be designed for “energy dissipation”. For buildings of
the most common materials in earthquake-resistant construction, namely concrete, steel or composite
(steel-concrete) buildings, two ductility classes (DCs) are foreseen for “dissipative” structures: Medium
(M) and High (H) ductility. DC M and H buildings are entitled values of the behavior factor q for force
reduction well above the value of q=1.5 (or 2 for steel or composite) attributed to overstrength. The global
energy dissipation and ductility capacity needed for values of q (well) above 1.5 is ensured via:
- measures to control the inelastic response mechanism, so that concentration of inelastic deformation
in a small part of the structure (mainly a soft story mechanism) and brittle failure modes are avoided;
- detailing of the plastic hinge regions for the inelastic deformations expected to develop there under
the design seismic action (the one for which the no-local-collapse requirement applies).
Concentration of inelastic deformations and soft story mechanisms are avoided by configuring and
proportioning the lateral-force resisting system so that vertical members remain practically straight – i.e.
elastic – above their base. Concrete wall or dual systems and braced steel or composite frames are
promoted and are capacity-designed to ensure that yielding takes place only at base of walls, only in the
tension diagonals of concentric braced frames, or in the special “seismic links” of eccentric braced frames.
In concrete, steel or composite moment frames, columns are capacity-designed to be stronger than the
beams, with an overstrength factor of 1.3 on beam design flexural capacities in their comparison with that
of columns. Concrete beams, columns and walls are capacity-designed against (brittle) shear failure.
Once the global inelastic response mechanism is controlled, the behavior factor q for reduction of elastic
forces is related to the global displacement ductility factor, µ, via the Vidic et al [3] q-µ-T relation:
µδ=q if T1≥TC, µδ=1+(q-1)TC/T1 if T1<TC (1)
where T1 = building fundamental period and TC = period at the upper limit of the constant acceleration
spectral region. µδ can ultimately be related to the local inelastic deformation demands in structural
members; e.g. for concrete members through the following approximation gives the local curvature
ductility factor, µφ: µφ=2µδ-1. Member detailing aims at providing the local deformation capacity to meet
these demands. For instance, the deformation capacity of flexural plastic hinges in concrete members is
controlled through the confining reinforcement in columns or in boundary elements of walls, and through
the compression reinforcement in beams, determined from µφ, which depends in turn on the q-factor
µφ=2q-1 if T1≥TC µφ=1+2(qo-1)TC/T1 if T1<TC (2)
Eq.(2) is based on Eq.(1) and on the approximation µφ=2µδ-1. So the spectrum of q-values gives rise to
detailing that also varies continuously. The provisions of EC8 for proportioning and detailing of concrete
members for energy dissipation in buildings of DC M and H are described in detail in Fardis [4].

The two upper Ductility Classes represent two different possible balances of strength and ductility, more
or less equivalent in terms of total material costs and achieved performance at the collapse prevention
level. DC M is slightly easier to design for and achieve at the site and may provide better performance in
moderate earthquakes. DC H may give better performance under motions (much) stronger than the design
seismic action. Unlike US codes, EC8 does not link selection of the ductility class to seismicity of the site
or the importance and occupancy of the building, nor puts any limit to their application. The choice is left
to the National Annex, which may in turn leave it to the designer, depending on the particular project.

Unless the Country objects through its National Annex, design directly on the basis of nonlinear analysis
is allowed, with members verified by comparing directly deformation supplies to demands. The definition
of acceptable member deformation limits is left to National Annexes, in the form of supplementary non-
contradictory information. To ensure a minimum global and local ductility for the so-designed buildings,
EC8 requires that they meet all DC M rules (i.e. those for member detailing, strong columns-weak beams
in frames, capacity design of concrete members in shear, etc.). By allowing design directly on the basis of
nonlinear analysis with member verification on the basis of deformations, the 1st generation of EN-
Eurocode 8 paves the way for fully displacement- and deformation-based design in the 2nd generation.
This approach is rapidly gaining acceptance in Europe, along with nonlinear analysis for the evaluation.

In the 2nd generation of EN-Eurocodes, the Section of Part 1 of EC8 on concrete buildings may see
extension of its scope to cover prestressing in members with plastic hinging and energy dissipation, as
well abandoning the curvature ductility factor as the link between global inelastic deformation demands
(cf. Eq. (2)) and member detailing for ductility, in favor of more physically meaningful approaches.

Analysis procedures for the design of new buildings

Part 1 of EC8 includes the following analysis options for building design or assessment of performance:
- Linear static (termed “lateral force” method).
- Linear modal response spectrum analysis.
- Nonlinear static analysis (“pushover”).
- Nonlinear dynamic (response time-history).
Unlike US codes, where linear static analysis is the reference, in EC8 the linear modal response spectrum
method is the standard procedure, applicable to all types of buildings. Countries are allowed to limit the
use of nonlinear analysis methods through the National Annex.
The lateral force procedure may be applied if the effects of higher modes are not significant, i.e. only if:
- In both horizontal directions the fundamental period T1 is less than 2 sec and 4 times the transition
period TC between the constant-acceleration and the constant-pseudovelocity regions of the spectrum.
- There are no significant irregularities in elevation.
Regarding irregularity, unlike US codes [5], [6], which set quantitative – albeit arbitrary – criteria based
on the heightwise distribution of story mass, stiffness and strength, EC8 introduces qualitative criteria,
easy to check at the preliminary design stage prior to any analysis:
- Lateral force-resisting systems continuous to the top of the – relevant part of the – building.
- Story mass and stiffness that is constant or reduces gradually and smoothly to the top.
- Individual setbacks less than 20% of the underlying storey if they are provided symmetrically on both
sides of the building, or 10% if they are unsymmetric, etc.
- – In frame buildings – smooth variation of story overstrength with respect to the design story shear.

In the lateral force method of analysis the base shear is computed by assigning the total mass to the 1st
translational mode in the horizontal direction of interest. It may use values of the fundamental period
estimated through empirical expressions – the same as in the SEAOC ’99 requirements (e.g. T1=0.085H3/4
for steel moment frames, T1=0.075H3/4 for RC frames or eccentric braced steel frames, T1=0.05H3/4 for all
other systems, with H = height from base in m). Such expressions represent lower (mean minus one
standard deviation) bounds to values inferred from measurements on buildings in California in past
earthquakes and give low values of the period, especially for buildings with lower required earthquake
resistance and hence less stiff. Use of a period calculated from mechanics – e.g. via the Rayleigh quotient
– is encouraged (as preferred in European design practice), without a check or a limitation of the value
with respect to the empirically calculated one. Moreover, if the fundamental period is less than twice the
transition period TC of the spectrum, in buildings with more than two stories the total lateral force is
reduced by 15% - to account for the difference between the 1st mode mass and the total. The objective is to
bring the results of the lateral force method closer to those of modal response spectrum analysis, and not
the other way around as in US codes. The total lateral force is distributed to the stories following a 1st
mode pattern of response accelerations, which may be taken as inverted triangular.

In the response spectrum analysis modal results are combined via rigorous application of the SRSS or
CQC rule, i.e. at the level of the final seismic action effects of interest (displacements, internal forces,
etc.). This is unlike the US approach, where modal response spectrum analysis is cast to look like linear
static, in that story lateral forces are calculated for each mode (from modal story accelerations) and
combined via the SRSS (or CQC) rule, for the structural system to be then analysed under the resulting
lateral forces as in the linear static procedure. A drawback of the rigorous EC8 approach is that when the
SRSS or CQC rule is applied to modal internal forces, signs are lost and, with them, physical meaning too.

Pushover analyses should be performed under two lateral load patterns: one corresponding to uniform
lateral accelerations and another similar to the lateral forces used in linear static (lateral force) analysis, if
applicable, or derived from a modal response spectrum one. The target displacement for pushover analysis
is derived via the N2 procedure by Fajfar [7], given in an informative annex. The key elements in this
procedure are: (a) the fitting of an elastic-perfectly plastic curve to the base shear-top displacement curve
of the building; (b) establishment of an equivalent SDOF system on the basis of a lateral displacement
pattern consistent with the postulated lateral load pattern and the calculation of its elastic stiffness as that
of the elastic-perfectly plastic curve; and (c) estimation of the target displacement from the 5%-damped
elastic spectrum, through the equal displacement rule, as modified for T1<TC via Eq.(1).

Nonlinear response-history analysis should use as input at least 3 artificial, recorded, or simulated records
(or 3 pairs of different records, for analysis in 3D), the mean elastic spectrum of which should no-where
fall below that of the design seismic action by more than 10%. Results are averaged, if at least 7 such
analyses are performed; otherwise the most unfavourable ones among the analyses performed are used.
In planwise regular buildings the analysis for the two horizontal components of the seismic action two
may use independent 2D models. Before the analysis a building can be characterized as regular-in-plan if:
- it has rigid diaphragms, nearly rectangular in plan (re-entrant corners reducing floor area by not more
than 5% each), with aspect ratio less than 4.
- the eccentricity between the story centers of mass and stiffness is less than 30% of the corresponding
story torsional radius (square root of ratio of torsional to lateral stiffness, with stiffness parameters
estimated in most cases from the moments of inertia of vertical elements), and
- at each story the torsional radius in each direction is less than the radius of gyration of the floor mass.
Two separate 2D models may also be used for buildings of ordinary importance with: (a) height less than
10m or 40% of the plan dimensions; (b) story centers of mass and stiffness approximately on (two)
vertical lines, and (c) in both horizontal directions torsional radius not less than the SRSS of the radius of
gyration of the floor in plan and of the projection of the eccentricity between centers of mass and stiffness
in that direction. If conditions (a) and (b) are fulfilled, but not condition (c), then two separate 2D models
may still be used, provided that all seismic action effects from the 2D analyses are increased by 25%.

Regardless of whether they are computed via a single 3D or two separate 2D models, seismic action
effects due to the individual horizontal components are combined through the SRSS or the 1:0.3 rule.
Maximum values of action effects estimated individually through the SRSS rule may be conservatively
assumed to take place at the same time. Nonetheless, more accurate and less conservative rules may be
introduced in the National Annex for the estimation of the likely simultaneous values of action effects due
to the different components of the seismic action. In buildings which are regular in plan and have
independent lateral-force-resisting systems in the two horizontal directions consisting solely of walls or
bracing systems, the effects of the two horizontal components do not need to be combined.

Accidental eccentricity is taken equal to 5% of the perpendicular plan dimension, without amplification
due to torsional-lateral coupling. Its effects may be calculated statically, by applying story torsional
moments to a 3D structural model, even when the modal response spectrum method is used in the analysis
of the response to the two horizontal components. The effects of accidental eccentricity may be accounted
for in a simpler and conservative way by amplifying the results of the – linear static or dynamic, or
nonlinear static – analysis for each horizontal component by 1+0.6x/L, where x denotes distance of the
element of interest from the center in plan and L the plan dimension, both normal to the direction of the
seismic action. This factor is derived assuming that torsional effects are fully resisted by the lateral-force-
resisting-elements in the horizontal direction considered and that such elements are uniformly distributed
in plan. In planwise regular buildings analysed with two separate 2D models, factor 0.6 is replaced by 1.2.

Unlike what happens in some US codes, amplification of eccentricities between centers of mass and
stiffness is not required. This is convenient, as in most cases story stiffness centers cannot be uniquely
defined; moreover, determination of their position at a level of accuracy and sophistication consistent with
the dynamic amplification of natural eccentricity, requires tedious additional analyses.

In all types of analysis the model should include only elements considered as part of the lateral-force-
resisting-system, termed “primary seismic elements”. “Secondary seismic elements” should collectively
account for less than 15% of the total lateral stiffness and should be distributed regularly and uniformly in
plan and elevation, so that they don’t affect the regularity classification of the structural system.

In linear analysis the elastic stiffness should be the secant stiffness to yielding, which for concrete and
masonry may be taken as half of the uncracked stiffness of the gross section. Nonlinear analysis may use
this value as pre-yield stiffness (pre- and post cracking stiffnesses may be considered also, if so-desired),
and may neglect the effect of strain hardening on post-yield stiffness (post-yield stiffness should be taken
negative, if there is significant strength degradation). Hysteresis models for nonlinear response-history
analysis should reflect realistically the energy dissipation within the expected range of displacements.
Nonlinear element models should use mean material properties, which are higher than nominal values.

For linear analysis displacements are calculated in general on the basis of the equal displacement rule, i.e.
by multiplying linear analysis results by the q-factor. Allowance is made for more accurate calculation,
including the q-µ-T relation of Eq.(1). Story drifts determined in this way are used to estimate the ratio of
P-∆ effects to 1st-order effects. If the ratio θ=Pδ/Vh exceeds 0.1 in any story, 1st-order analysis results are
divided by 1-θ. P-∆ effects are calculated on the basis of the secant-to-peak-drift story stiffness, instead of
the elastic stiffness used in US codes, influencing analysis results more often and to a larger extent.

Non-engineered masonry infills producing irregularities in plan or elevation should be taken into account
in the analysis of steel or composite moment frames of DC H, as well as in DC H concrete buildings with
walls providing less than 50% of lateral force resistance. Normally, irregular distribution of infills in plan
may be considered by doubling the effects of accidental eccentricity. Very irregular arrangements, as e.g.
when infills are concentrated mainly on one or two sides of the perimeter, require instead a 3D model that
explicitly includes the infills, and a sensitivity analysis of the effect of their properties and position
(including removal of one out of 3 or 4 infill panels from the model). Attention is called to the possible
effects on structural elements furthest away from the sides of the plan where infills may be concentrated.
A reduction in the amount of infills relative to the story above (as in open ground-stories) should be taken
into account by amplifying seismic action effects from the analysis so that the deficit in infill shear
strength in the story is compensated by the increase in resistance of vertical members of the frame there.

The current trend in Europe, is towards nonlinear analysis methods: the traditional response-history
version and the more fashionable pushover analysis. In the draft Part 3 of EC8 pushover analysis is the
reference method for the assessment of existing buildings and the evaluation of their retrofitting. If the
future is the continuation of the past and present, in the 2nd generation of EC8 nonlinear analysis methods
will assume greater importance for the design of new buildings as well. This evolution will depend not so
much on developments in nonlinear models for members under monotonic or cyclic loading, but on the
emergence and wide acceptance of member verification criteria based on deformations. To be established
as the prime workhorse for seismic design of new concrete buildings, nonlinear analysis will have to
overcome a difficulty specific to concrete: that key parameters of nonlinear member models (e.g. the pre-
yield stiffness and the yield moment) depend on the amount of reinforcement, which is not known prior to
detailed design. Ad hoc procedures may be used to overcome this difficulty: e.g. beam reinforcement may
be first proportioned for gravity loads, that of columns may then be obtained from the strong column-weak
beam rule, accounting for minimum reinforcement, to provide an estimate of member pre-yield stiffness
and yield moments for the (first round of) nonlinear analysis, that will give the deformation demands for
which members will be verified and their transverse reinforcement proportioned and detailed.

Another point where there is room for improvement of the present EC8 rules is in the analysis methods
and approximations for the inelastic torsional response due to natural or accidental eccentricities in plan.

Behavior factor q for reduction of the forces from elastic analysis

In structures designed for energy dissipation the behavior factor, q, that reduces the forces in the elastic
structure, is linked, directly or indirectly, to the ductility and deformation demands on members and
connections, and hence to the corresponding detailing requirements. The value of q depends on the
structural material, the type of lateral-force-resisting-system and the ductility class selected for the design.
The values quoted here are the ones adopted by EC8 for the most common structural materials in
earthquake resistant construction, namely for concrete, steel and composite (steel-concrete) buildings.
Overstrength due to materials is considered to account for a q-factor of 1.5 (or up to 2 for steel or
composite buildings), even in buildings without energy dissipation. In buildings with energy dissipation
and structural system redundancy the q-factor incorporates a factor αu/α1 (denoted here for convenience by
αR) that accounts for (structural) system overstrength. The αR factor is the ratio of the seismic action at
development of a full plastic mechanism (i.e. of a fully yielded structure) to the one that causes formation
of the first plastic hinge in the system – in both cases with the gravity loads acting simultaneously with the
seismic action. Default values of αR in buildings regular in plan (defined in the previous Section) are:
- 1.0 for: concrete wall systems with just two uncoupled walls per direction;
- 1.1 for: (a) one-story moment frames, (b) wall concrete systems with more than two uncoupled walls
per direction, (c) steel or composite frames with concrete infills, (d) wall concrete systems with
encased steel sections as wall boundary elements, (e) systems of concrete walls - possibly
with encased steel sections as boundary elements - coupled by steel or composite beams;
- 1.2 for: (a) one-bay multistory moment frames, (b) dual concrete systems with walls providing 50% to
65% of lateral force resistance, (c) coupled wall concrete systems, (d) eccentric-braced steel or
composite frames, (e) steel or composite systems consisting of concentric-braced and moment frames;
- 1.3 for multistory multi-bay frames, (f) systems of walls with steel or composite boundary elements
and web consisting of a steel plate with concrete overlay on one or both sides.
In buildings that are not regular in plan, the default αR-value is the average of 1 and the default value
listed above. Values higher than the default may be used (up to a maximum of 1.5 for concrete, or of 1.6
for steel and composite), if they are confirmed through a pushover analysis of the so-designed structure.
Table 1. q-factors for structural systems regular in elevation
Material Lateral-load resisting structural system DCM DCH
steel or inverted pendulum systems (≥50% of mass in upper third of height, or with energy 2 2αR
composite dissipation at base of single element)
steel or moment-resisting frames, or eccentrically-braced frames 4 5αR
composite
steel or frames with concentric diagonal or X-bracings (with only the tension diagonals 4 4
composite considered for seismic resistance)
steel or frames with concentric V-braces (with both tension and compression diagonals 2 2.5
composite considered for seismic resistance):
steel or moment-resisting frames combined with concentric (diagonal, V or X) bracings 4 4αR
composite
steel or moment-resisting frames with concrete or masonry infills in contact with frame, 2 2
composite without positive connection
composite 1) steel or composite frames with connected concrete infills; 2) concrete walls with 3αR 4αR
encased steel sections as boundary elements at the edges; or 3) walls with steel or
composite boundary elements and web of a steel plate with concrete overlay(s)
composite 1) concrete walls of the type (2) above; or 2) concrete walls coupled through steel 3αR 4.5αR
or composite beams
concrete inverted pendulum systems (defined above for steel or composite) 1.5 2
concrete torsionally sensitive (radius of gyration of floor mass > torsional radius in one or 2 3
both directions: i.e. lateral-load-resisting system concentrated near plan center)
concrete system of walls with resistance ≥65% of seismic base shear and with >50% of wall 3 4αR
resistance from uncoupled walls (: coupling reduces wall base moments by <25%)
concrete all concrete systems other than above 3αR 4.5αR
For concrete, steel or composite buildings that are regular in elevation, the q-factor for Ductility Class M
or H is given in Table 1. This is the “basic” value, qo, of the q-factor, used to estimate the ductility
demands in element regions intended for energy dissipation and detailed accordingly. In buildings with
vertical irregularities, seismic action effects from linear analysis are calculated from a q-factor equal to
80% of the “basic” value, qo, but not less than the value of 1.5 for concrete, or of 2 for steel or composite,
which is considered available due to overstrength alone. A building may have different q-factors in the
two main horizontal directions, depending on the structural system and its vertical regularity classification
in these directions, but not due to Ductility Class, which is the same for the whole building.

Design of the foundation and of its elements

Unlike US codes [5], [6], which allow reduction of overturning moment at the base due to uplift by 25%
for linear static analysis or by 10% for a response spectrum one, in DC M and H buildings verification of
the foundation soil and proportioning of foundation elements are also based on seismic action effects
derived from capacity design. This is due to the importance of the foundation for the integrity of the whole
and the difficulty to access, inspect and repair damaged foundations. Capacity design seismic effects, to
be superimposed to those due to gravity loads, are the – factored by 1.2 – seismic action effects which
occur at the intended plastification of the component of the superstructure with the highest influence on
the foundation element in question (i.e. the base of a frame column or wall in the direction which yields
first in bending). For common foundations of more than one vertical element (e.g. rafts, foundation beams,
etc.) plastification in the element with the largest design seismic shear is considered in these capacity-
design calculations. Foundation elements (including piles) proportioned for the above capacity-design
load effects, follow the detailing etc. rules of DC L structures, i.e. design according to Eurocode 2 alone.
Alternatively, they may be proportioned for seismic action effects derived from the analysis; then they
should meet all the proportioning and detailing rules of the corresponding DC for members of the
superstructure (e.g. capacity design in shear for foundation- or tie-beams).

THE DRAFT EN-EUROCODE 8 FOR SEISMIC REHABILITATION OF BUILDINGS

Levels of Performance (Limit States) and Hazard Levels (Seismic action)

Part 3 of EC8 [2] adopts a fully performance-based approach for the evaluation and rehabilitation of
existing buildings; three performance levels (termed in the European tradition “Limit States”) are defined:
- “Near Collapse” (NC), similar to the “Collapse Prevention” level in US codes [7], [8]. The structure
is heavily damaged, may have large permanent drifts, retains little residual strength and stiffness
against lateral loads, but vertical elements can still carry the gravity loads. Most non-structural
elements have collapsed. The building may not survive another earthquake, not even a strong
aftershock. Repair may not be technically feasible and is not economically justified.
- The “Significant Damage” (SD) level, corresponding to the “Life Safety” level in the US [7], [8]and
to the local-collapse prevention performance level for which new structures are designed according to
Part 1 of EC8. The structure is significantly damaged, may have moderate permanent drifts, but
retains some residual lateral strength and stiffness and certainly its vertical load-bearing capacity.
Non-structural components are damaged, but have not collapsed. Repair may be uneconomic.
- The “Damage Limitation” (DL) level, corresponding to the “Immediate Occupancy” level in the USA
[7], [8]. The structure is only lightly damaged, has no permanent drifts and its elements have no
permanent deformations, retain fully their strength and stiffness and do not need repair. Non-
structural components may have distributed cracking and can be easily and economically repaired.
To allow national determination in all safety-related issues, the “Seismic Hazard” levels for which the
three “Limit States” will be required are NDPs. For ordinary buildings, the recommendation in EC8 is for
a 225-year earthquake (20% exceedance probability in 50 years), a 475-year event (10% in 50 years) and a
2475-year one (2% in 50 years), for the DL, the SD and the NC “Limit State”, respectively.
Performance differentiation of essential or large occupancy buildings from ordinary ones is effected as in
new buildings: by multiplying the seismic action with the “importance factor”, which has the
recommended values given in Part 1. The same spectral shape applies at all “hazard levels”. It is
noteworthy that as the SD “Limit State” conceptually corresponds to the local-collapse prevention level
for which new structures are designed according to Part 1 of EC8, Parts 3 and 1 recommend the same
seismic action for both, without differentiation of the “Performance Objective” between new and existing
buildings. Any such differentiation comes through the “compliance criteria” specified in Parts 1 and 3 for
this single performance level. National Authorities are free to relax this requirement for existing buildings
(by specifying in the National Annex a lower seismic action for the SD “Limit State”), to reduce the
number of buildings that need retrofitting and/or make retrofitting more attractive from the economic
point of view. National Authorities will also specify whether all three “Limit States” will have to be
verified under the corresponding “Seismic Hazard” level, or whether verification of one or two “Limit
States” for the corresponding “Seismic Hazard” level may suffice.

Analysis methods for the calculation of deformation demands and criteria of applicability
Part 3 of EC8 has adopted a fully displacement-based approach. So the objective of the analysis for the
seismic action of interest is the estimation of deformation demands in structural members.

Part 3 provides the same four analysis options of Part 1 of EC8. In all analysis methods the seismic action
is given by the 5%-damped elastic response spectrum or the quantities of interest derived from it: the
target displacement for nonlinear static analysis, or acceleration time-histories for nonlinear dynamic
analysis. The spectrum is anchored to the PGA corresponding to the “hazard level” chosen for the “Limit
State” for which analysis results will be used, with multiplication by the “importance factor” appropriate
for the building. All types of analysis, except the nonlinear dynamic, employ the equal-displacement rule,
at the level of member deformations, e.g. chord-rotation demands for the first two types of analysis, or of
the displacement of an equivalent SDOF system for pushover analysis.

If linear analysis is applied, internal forces in “brittle” members are estimated as in “capacity-design”, i.e.
from equilibrium assuming that “ductile” locations delivering force to them develop their force capacity,
or the force demand from the analysis, whichever is less. Force capacities are estimated from expected
values of material strengths derived from the available information, multiplied by a “confidence factor”
greater than 1 that depends on the amount and reliability of information on the as-built structure.

The criteria for applicability of linear-elastic analysis, static or dynamic, are the following:
- In all “brittle” elements force capacity exceeds demand estimated according to the paragraph above.
- There is a fairly uniform distribution of ductility demands over the entire structure. If the ratio D/C of
bending moment at the end of a member from elastic analysis to the corresponding flexural capacity
is taken as a measure of the local ductility demand (this ratio is roughly equal to the demand value of
the chord-rotation ductility ratio), then the maximum value of the D/C-ratio over the entire structure
should not exceed the minimum value of the D/C-ratio over the structure by a factor between 2 to 3.
the maximum permissible value of the D/C-ratio within this range is a NDP with a recommended
value of 2. In this criterion D/C values less than 1 are taken equal to 1, indicating elastic response.
If both these conditions are met, the relevant criteria of Part 1 of EC8 determine whether linear static
analysis (“lateral force” procedure) may be used instead of the modal response spectrum method. No limit
is set for the absolute value of the ductility demands of “ductile” elements for linear analysis to be
applicable. Nonetheless, the limitation of the ratio of maximum to minimum value of the D/C-ratio below
a ceiling between 2 and 3 (with a recommended value of 2) poses a very severe restriction to the
applicability of linear-elastic analysis (except at the DL “Limit State”, where elastic response is required).

Each one of the four analysis methods should follow the rules specified in Part 1 for new buildings.
According to these rules the elastic stiffness should be the secant stiffness to yielding, which Part 1 allows
taking equal to half of the uncracked stiffness of the gross section in concrete or masonry buildings. This
rule is meant to provide safe-sided estimates of force demands in force-based design; nonetheless, it
underestimates the effective secant-to-yield rigidity of members and provides unconservative estimates of
chord rotation demands for verification in displacement-based evaluation and rehabilitation, as in Part 3 of
EC8. To avoid this, the secant-to-yield rigidity of concrete members may be determined from the yield
moment and the chord rotation at yielding, estimated through semi-empirical expressions given in the
informative Annex of Part 3 devoted to concrete buildings.

Compliance criteria for the three “Limit States”

Following Part 1 of EC8, Part 3 distinguishes the elements considered as part of the lateral-force-resisting
system (“primary”) from those which are not (“secondary”). The restriction in Part 1 that “secondary”
elements should collectively account for not more than 15% of the total lateral stiffness is waived. As
more relaxed compliance criteria apply for “secondary” elements, the engineer is free to designate
elements of the existing or of the retrofitted building as “secondary”, depending on the outcome of rounds
of analysis and verifications with alternative trial designations of some elements as “primary” or as
“secondary”. What the engineer is not allowed to do is to deliberately choose the distribution of
“secondary” elements in plan and elevation so that the regularity classification of the structural system
changes from irregular to regular (as this may affect the method of analysis allowed to be applied).

A distinction is also made in Part 3 between “ductile” and “brittle” elements. In concrete buildings
members are characterized as “ductile” in bending and as “brittle” in shear. Compliance criteria of
“ductile” elements are in terms of deformations, while those of “brittle” elements are in terms of forces:
- At the NC “Limit State” “ductile” elements are allowed to reach their (expected) ultimate deformation
capacity. The Annex on concrete buildings specifies this as the ultimate chord rotation capacity,
computed from mean material strengths divided by a “confidence factor” that depends on the amount
and reliability of the information available for the as-built structure; moreover an overall safety factor
of 1.5 is applied to the ultimate chord rotation capacity of “primary” elements. Force demands on
“brittle” elements should remain below their ultimate force resistance, computed from mean material
strengths. For primary concrete elements these mean material strengths are divided by the material
partial factors times the “confidence factor” that depends on the amount and reliability of
information; moreover, the so-computed shear resistance is further divided by a safety factor of 1.15.
For secondary elements, factoring of mean material strengths by material partial factors or
“confidence factors”, or of the shear resistance by an additional safety factor, is not required.
- At the SD “Limit State” deformations (chord rotations at member ends) of “ductile” elements are
limited to a fraction of their ultimate deformation capacity. The Annex for concrete buildings
specifies this as 75% of the ultimate chord rotation capacity, with the additional safety factor of 1.5
for “primary” elements. In the force-based verification of “brittle” elements the capacity is taken the
same as in the NC “Limit State” and as the demand is also (essentially) the same, verification at the
SD “Limit State” may be omitted as redundant, if already done at the NC “Limit State”.
- At the “Damage Limitation” (DL) level, “ductile” and “brittle”, “primary” or “secondary” elements
alike, are required to remain elastic (below yielding).

The informative Annex on concrete buildings provides empirical or semi-empirical expressions for the
ultimate chord rotation of concrete members, as affected by their geometry, the amount and detailing of
(longitudinal and transverse) reinforcement, the presence and length of lap splices, etc. It gives also semi-
empirical expressions for their shear resistance, as controlled by yielding of stirrups or (for “short
columns”, i.e. those with shear-span-ratio less than 2, and for walls) by diagonal web crushing, and as it
decreases with the magnitude of post-yield cyclic deformations. Expressions and design rules are also
given for the enhancement of member ultimate chord rotation and shear resistance effected by jackets of
concrete, steel or fiber-reinforced polymers (FRPs).

Prospects for the evolution of Part 3 of Eurocode 8 towards the 2nd generation of EN-Eurocodes
Unlike Part 1 of the EN-Eurocode 8, which has evolved from the ENV version of 1994 with extensive
changes in rules, addition of many new provisions and significant enlargement of the scope, but without
radical changes in the overall philosophy, the draft Part 3 represents a revolutionary change over its 1996
predecessor. Therefore, technical developments towards the 2nd generation of Part 3 of EC8 will depend
on how this novel 1st generation EN-Eurocode will be received and applied in European seismic
rehabilitation practice. As another difference with Part 1 - which consists almost exclusively of normative
text or Annexes albeit with a few dozens of NDPs - Part 3 has a brief and general normative part with very
few NDPs, but three long and very detailed informative Annexes - one for each of the three main
materials: concrete, steel or composite and masonry. As Member States may accept informative Annexes
as source of information or replace them with other (national) alternatives, the form and the technical
contents of the 2nd generation of Part 3 of EC8 will strongly depend on how the three Annexes will be
received in Member States. If they are accepted as source of information in practically all Member States
publishing a National Annex for the EN-Eurocode 8, it is very likely that the European Commission will
enrich the limited normative part of Part 3 of EC8 by upgrading (some of) the three Annexes to
Normative, even before launching the 2nd generation of EN-Eurocodes.

ACKNOWLEDGMENTS

Part 1 of EC8 is the outcome of the work of several individuals in the CEN body in charge of Eurocode 8,
namely CEN/TC250/SC8, and in its project teams PT1 and PT2 led by Profs. C. Oliveira and J.
Bouwkamp, respectively. The draft of Part 3 was developed by a small project team, PT5, led by Prof. P.
Pinto. The technical and administrative contribution of the Technical Secretary of CEN/TC250/SC8, Dr.
E. Cansado Carvalho, deserves a special acknowledgment.

REFERENCES

1. CEN, “European Standard EN 1998-1: Eurocode 8: Design of structures for earthquake resistance.
Part 1: General rules, seismic actions and rules for buildings”. Stage 51 Draft, Brusells: Comite
Europeen de Normalisation, December 2003.
2. CEN. “Draft European Standard prEN 1998-3: Eurocode 8: Design of structures for earthquake
resistance. Part 3: Assessment and retrofitting of buildings”. Draft No.5. Doc CEN/TC250/SC8/N388,
Brusells: Comite Europeen de Normalisation, January 2004.
3. Vidic, T, Fajfar, P, Fischinger, M. “Consistent inelastic design spectra: strength and displacement”.
Earthquake Engineering and Structural Dynamics 1994; 23: 502-521.
4. Fardis, MN. “European Developments in Codified Seismic Design of Concrete Structures”,
Proceedings of 1st fib Congress, (Keynote paper K-20), Osaka. October 2002.
5. BSSC. “NEHRP recommended provisions for seismic regulations for new buildings and other
structures”. 2000 Edition, Washington, D.C.: Building Seismic Safety Council, March 2001.
6. SEAOC. “Recommended lateral force requirements and commentary”. Sacramento, CA: Seismology
Committee, Structural Engineers Association of California, 1999.
7. Fajfar, P. “A non-linear analysis method for performance-based seismic design”. Earthquake Spectra
2000; 16(3): 573-593.
8. ASCE. Prestandard for the seismic rehabilitation of buildings. Reston, VA: American Society of Civil
Engineers (FEMA Report 356), 2000.